# Amazon SEO Guide: Title Rules, Keywords and Funnel Fixes

> Amazon SEO guide: how Amazon matches and ranks listings, the 75-character title cap and twice-per-word rule, plus a funnel check for Brand Registry sellers.

- URL: https://missiongrowth.io/blog/amazon-seo
- Published: 2026-10-10
- Author: Furkan Aktaş, Co-Founder, Mission Growth
- Publisher: Mission Growth. Company facts: https://missiongrowth.io/llms.txt

Amazon SEO is the work of getting your listing matched to a shopper's search and then bought once it shows up. Most checklists hand you a dozen things to optimize and no way to know which one is holding you back.

This guide gives you the rules Amazon itself publishes for titles and repeated words, and a three-ratio check that names the stage of the funnel where your listing loses the sale. If you have a live listing, you can run it this week.

In this guide:

- How Amazon matches and then ranks products, and what is only seller inference
- What each listing field is for, with Amazon's own limits
- What Amazon enforces on keyword stuffing, with a worked audit
- How to find the weak funnel stage with Search Query Performance
- What Amazon SEO costs and which certification exists

## Amazon SEO: how Amazon matches and ranks listings

Amazon SEO is the work of making a product listing easy for Amazon to match to a search and easy for shoppers to buy once it appears.

Amazon search engine optimization is the same thing under its long name. Put simply, that is what Amazon SEO is: two jobs, in order. The first job is to be matched. The second is to convert, because what shoppers do after they see you feeds back into where you appear.

Shoppers on Amazon arrive with purchase intent. The results page lets customers compare prices, shipping options and more, and Amazon's models learn from what they then buy. If you also run your own store, [ecommerce seo](https://missiongrowth.io/blog/ecommerce-seo) covers that side; a marketplace listing is a different system with a different feedback loop.

Amazon's researchers described how product search works in a [2019 paper on semantic product search](https://arxiv.org/abs/1907.00937). Search runs in two stages, matching and then ranking:

- **Matching.** Products that contain the query words are the primary candidates. Products with prior behavioral associations, meaning items bought or clicked after that query, join the candidate set too.
- **Ranking.** The candidates are ordered by a machine-learned rank function that optimizes for customer satisfaction and business metrics.

The same paper describes a network trained on purchases and clicks that complements the word-matching engine, and it separates purchased items from items that were impressed but not purchased. The reason is that pure lexical matching falls short on synonyms and spelling variants such as "woman" versus "women".

::figure{src="/blog/figures/amazon-seo-1.svg" alt="Flow chain of Amazon SEO in two stages: listing words and behavior linked products form the candidate set, then a learned rank orders it." caption="Amazon SEO has two stages: be matched to the query, then be ranked on what shoppers do." width="720" height="216"}

What follows for you: your words are the main way in, behavior can add you to the candidate set, and behavior always orders you once you are in. Both jobs matter, which is why a keyword-perfect listing with a weak image still stalls.

One limit applies. This is a 2019 research system tested on a large e-commerce website, so treat it as the logic of the approach and not a statement of today's production ranker.

Sellers call the system Amazon A9 and, later, A10; Amazon's seller guide names neither. A9 and A10 are seller shorthand for the ranker: sellers call it whichever version is current, so read "Amazon's A10 update" as a seller's label.

The same caution covers every named weight. Statements such as "conversion rate is a direct ranking factor" or "sales velocity counts for more than ads" are seller inference. The split looks like this:

- **What Amazon documents:** a search box, filters, a results page, sales rank and ads, listed without weights.
- **What sellers infer:** an ordering of conversion, sales velocity, click-through rate and reviews, each with a named weight.

Behavior data feeds matching as well as ranking, so no single factor can be pulled out and weighed. Treat each inferred factor as something you can measure on your own listing, and the funnel check below does that.

### Rufus and the question of whether Amazon SEO is dead

Amazon SEO is not dead: Amazon's Rufus shopping assistant was used by 300 million+ customers in 2025, according to [Amazon's fourth-quarter 2025 results](https://ir.aboutamazon.com/news-release/news-release-details/2026/Amazon-com-Announces-Fourth-Quarter-Results/).

That release gives usage and sales figures. It does not say which listing fields Rufus draws on, so any field-level Rufus advice would be a guess.

What we can say is that matching and ranking still run on search, and Amazon still reports a search funnel for your products (the check in this guide uses it). A listing that is matched, clicked and bought is the same asset either way.

For the Google-wide version of that question, the guide on [is seo dead because of ai](https://missiongrowth.io/blog/will-ai-replace-seo) works through it.

If your products also surface in AI shopping answers outside Amazon, [how to optimize for chatgpt shopping](https://missiongrowth.io/blog/chatgpt-shopping-optimization) covers the product data side.

## Amazon product listing optimization, field by field

Amazon product listing optimization means giving every field one job: titles and backend terms are how you get matched, and the main image, bullets and A+ Content are how you win the click and the sale.

A keyword that appears in every field does not work harder. Optimizing Amazon listings well means deciding which field carries each phrase, then leaving it alone.

### Keyword data by situation

Amazon SEO keywords come from different places depending on what you have access to. Brand Registry is the gate, and Amazon describes it as a free resource.

| Your situation | Start from | Limit |
|---|---|---|
| New product, no Brand Registry | Amazon autocomplete and manual searches | Shows the phrases shoppers type, not how your listing performs |
| Registered brand | Top Search Terms dashboard | Needs Brand Registry |
| Live listing with impressions | Search Query Performance | Needs Brand Registry; the API report also needs the Brand Analytics role |
| Any situation | Third-party tools | Exist; this guide does not compare them |

Without Brand Registry, start from Amazon autocomplete and type your product the way a buyer would. With Brand Registry, Top Search Terms shows what shoppers search for, and Search Query Performance shows how your own ASINs fare on those queries. That second report drives the funnel check in the last section.

### Product titles

Since July 27, 2026, Amazon's product title limit is 75 characters including spaces in every category except media. It was 200 characters for most categories before that date, and Amazon's seller guide recommends no more than 60. Item Highlights, a separate field of up to 125 characters, now carries materials and recommended use cases.

The 75 and the 60 answer different questions, and a third set floats around in practitioner advice:

::dataset{key="amazon-title-length-numbers" name="Amazon title length numbers and their sources"}

| Number | What it measures | Where it comes from |
|---|---|---|
| 75 characters | The enforced cap in every category except media, from July 27, 2026 | Amazon's June 10, 2026 Seller Central announcement |
| 200 characters | The earlier cap for most categories, in force from January 21, 2025 | Amazon's January 2025 title policy |
| No more than 60 | Amazon's advice for a readable title | Amazon's seller guide, which still prints the old 200 limit |
| Other length targets | Longer or shorter figures practitioners suggest | No Amazon source |

Lead with the main keyword and the product type inside the first 60 characters, and treat 75 as the ceiling. The advice to put "your top keywords in the first five words" is a practitioner heuristic that Amazon never published. It works as a side effect of front-loading: the phrase a shopper cares about shows up before the title gets truncated.

Two other rules from Amazon's January 2025 policy matter; check them against its current title requirements page. A title may not contain the same word more than twice (prepositions, articles and conjunctions are exempt), and the characters ! $ ? _ { } ^ ¬ ¦ are not allowed unless they are part of the brand name. The next section audits a title against the twice rule.

Google's title tags are a different field with their own limits; for that side, see [title tag length](https://missiongrowth.io/blog/title-tags).

### Bullets and descriptions

Amazon's own bullet format is a one or two word lead, then an explanation in 100 characters or fewer. That makes bullets the place where a shopper who has clicked decides whether to add to cart.

For descriptions, Amazon's rule of thumb is that a key phrase used once rarely needs a repeat. Use the description for the facts a bullet has no room for, not for a second pass at the title's keywords.

### Backend search terms

Backend search terms are not visible to customers, yet they can still lift a product on specific queries. Amazon says to include synonyms, abbreviations and alternative names for the product, and to avoid redundant or irrelevant terms.

That makes backend terms the home for the variants that would clutter a title: the abbreviation, the other name for the same product. Repeating a title word here adds nothing.

The field has a length limit, and Seller Central shows it where you type.

### Images and A+ Content

The main image carries the click. It sits next to your title and price on the results page, and it is the first thing a shopper compares.

A+ Content is free for brands enrolled in Brand Registry, and Amazon says it can help increase sales. This guide does not claim A+ text is indexed for search. Treat it as a cart-add lever: it sits on the detail page, where the shopper has already clicked.

## Keyword stuffing on Amazon: what gets enforced

Keyword stuffing on Amazon is repeating terms across a listing beyond what shoppers need, and Amazon sets a hard limit in titles and warns that stuffing hurts rankings everywhere else.

Amazon's seller guide warns that stuffing keywords creates a poor customer experience and hurts rankings. The written rules, field by field:

| Where | Rule | What happens |
|---|---|---|
| Title | 75 characters; the same word at most twice (January 2025 policy); no ! $ ? _ { } ^ ¬ ¦ | Non-compliant titles are flagged in Manage All Inventory |
| Existing titles | Brand owners get override suggestions in Review Listing Updates (2025) and AI-written title recommendations in Review Listings Changes (2026) | 14 days to act, then Amazon updates the title; listings remain active during the process |
| Description | A key phrase once is usually enough | Guidance in the seller guide |
| Backend terms | No redundant or irrelevant terms | Guidance in the seller guide |

The title rule is the only one with a count: the same word more than twice fails. That makes it the one you can audit today.

### A worked audit

The example listing below is invented for illustration. Its title runs 116 characters:

"Water Bottle Insulated Water Bottle Stainless Steel Water Bottle 32 oz Leakproof Bottle for Hiking Gym Sports Bottle"

Count each word. "Water" appears 3 times and "bottle" appears 5 times, against a limit of twice. The title fails on both words.

The rewrite is 68 characters, inside the 75-character cap, with no repeated word:

"AcmePro Insulated Stainless Steel Water Bottle, 32 oz, Leakproof Lid"

::figure{src="/blog/figures/amazon-seo-3.svg" alt="Side-by-side of an example stuffed Amazon title of 116 characters and its 68-character rewrite that uses each word once." caption="A stuffed 116-character title repeats water 3 times and bottle 5 times; the rewrite is 68 characters with no repeat." width="720" height="294"}

The rewrite keeps the searchable words and moves "for Hiking and Gym" to Item Highlights. "Bottle" once still matches the query word. The paper's point about variants like "woman" versus "women" is a reason to list the variant once in backend terms, not to repeat the word. Amazon's own rule of thumb for descriptions points the same way: say the phrase once.

If your titles predate the January 21, 2025 policy, check Manage All Inventory for non-compliant titles, starting with any over 75 characters, which Amazon has been rewriting gradually since July 27, 2026. Brand owners see override suggestions in Review Listing Updates and have 14 days to act before Amazon updates the title. Amazon states that listings remain active during this process.

## Find the weak stage with Search Query Performance

Search Query Performance shows, for each query, your ASIN's share of impressions, clicks, cart adds and purchases, and the stage where your share drops is the one to fix.

It is for Brand Registry brands. Amazon's Selling Partner API documentation describes the report as impressions, clicks, cart adds and purchases for a list of ASINs and a date range you specify, in weekly, monthly or quarterly periods. It also returns query totals and rates, plus median prices and shipping-speed counts at each stage.

The method is share carry-through. Divide each stage's share by the share before it. The lowest ratio is the leaky stage. Here is how to run it:

1. Pick a query you target.
2. Read your ASIN's share at each of the four stages.
3. Divide every share by the one before it, which gives three ratios.
4. The lowest ratio names the stage that loses the sale.
5. Compare your median click price and shipping-speed counts with the query totals in the same report.
6. Change one field that controls that stage.
7. Re-read the report next period and compare.

The ratios tell you which fields to look at:

| Ratio | What it tests | Fields that control it |
|---|---|---|
| Click share over impression share | Does the results-page listing earn the click? | Main image, title, price and delivery shown |
| Cart-add share over click share | Does the detail page convince? | Bullets, A+ Content, images |
| Purchase share over cart-add share | Does the offer close the sale? | Price, delivery, reviews (as inference) |

Price needs its own note. Amazon says adjusting your list price can improve ranking and sales, and the report gives you the median click price for your ASIN against the total, so you can see whether you sit far from the query's median.

Amazon documents no weight for price, so read it as a gap to close. For example, if your median click price sits well above the query's, test the price before you touch the title.

Test reviews as a conversion lever at the cart-add and purchase stages. Amazon documents no review weight, so treat them as inference, same as any named factor above.

### A worked example

Take an example listing, invented for illustration. For one query it has an impression share of 12%, a click share of 5%, a cart-add share of 4.5% and a purchase share of 4%.

Dividing a 5% click share by a 12% impression share gives a carry-through of 0.42. Cart-add share over click share is 4.5 over 5, or 0.90. Purchase share over cart-add share is 4 over 4.5, or 0.89.

::figure{src="/blog/figures/amazon-seo-2.svg" alt="Funnel of an example Amazon listing: 12% impression share, 5% click share, 4.5% cart-add share and 4% purchase share, with the click stage losing most." caption="In this example listing, share falls from 12% of impressions to 5% of clicks, so the click stage leaks." width="720" height="290"}

The lowest of the three is 0.42, so the click stage leaks. People see the listing and mostly pass. The shoppers who do click then convert at a steady 0.90 and 0.89, so the detail page is fine and rewriting bullets would waste the week.

The next edit comes from the first row of the stage table: the main image, the title wording or the price against the median click price. Change one of them, then re-read the report next period.

Zero impressions on a target query points back to matching. Go back to the title and backend terms for that phrase.

Without Brand Registry, you cannot run the ratios, because the report is gated. Use Amazon autocomplete and manual searches to see what shows up, and treat the method as what registering would give you. For deciding [which funnel stage to fix first](https://missiongrowth.io/blog/conversion-funnel-optimization) outside Amazon, the same logic holds on any store.

Your next step: open Search Query Performance, pick the one query that matters most, and compute the three ratios before you touch a field.

## Amazon SEO cost, certification and tools

Doing Amazon SEO yourself costs time plus whatever third-party tools you add, because the free Amazon tools cover the core keyword data and learning material.

Brand Registry is free, per Amazon. It unlocks Search Query Performance and Top Search Terms, and A+ Content is free for registered brands. That covers keyword data, diagnosis and enhanced content without a subscription.

Agency pricing is a separate question; [how much does seo cost](https://missiongrowth.io/blog/seo-cost) covers the general ranges. Third-party tools exist for Amazon keyword research and tracking. The free data above already answers the first decision, so start there.

Amazon's free certifications sit in Amazon Ads Academy, a free resource with courses, certifications, boot camps and videos that teach advertisers to use Amazon Ads products. They teach advertising products and say nothing about listing search. An "Amazon SEO" certificate comes from whoever sells the course, so treat it as a vendor credential.

Amazon needs your words to match and your listing to convert, so measure which one is failing instead of working through a checklist. Open your Search Query Performance report this week, divide the shares, and fix the one field that controls the lowest ratio.

## FAQ

### What does Amazon SEO stand for?

Amazon SEO stands for Amazon search engine optimization. It means improving your product listing so Amazon matches it to more relevant searches and shoppers buy it when it appears. The work covers keyword choice, titles, bullets, backend terms, images and the conversion that follows.

### Is Amazon SEO worth it now that AI shopping exists?

Yes. Amazon's Rufus shopping assistant was used by 300 million+ customers in 2025, but matching and ranking still run on search, and Amazon still reports a search funnel for your products. Amazon's release does not say which listing fields Rufus reads, so field-level Rufus advice would be a guess.

### How long should an Amazon title be?

Since July 27, 2026 the limit is 75 characters including spaces in every category except media (it was 200 for most categories before), and Amazon's seller guide recommends no more than 60. Lead with the main keyword and product type inside the first 60 characters, and use the rest only if shoppers need it. Under Amazon's January 2025 policy the same word may appear at most twice.

### Does price affect Amazon ranking?

Amazon says adjusting your list price can improve ranking and sales, but it states no weight. Search Query Performance reports median prices at each funnel stage, so you can compare your price with the query's median before you change it.

### Does Amazon favor FBA sellers in search?

No source in this guide documents an FBA weight in search. What Amazon does report is shipping-speed counts at each funnel stage in Search Query Performance, so you can see how delivery speed compares on a query before you assume it matters.

### How do I track Amazon SEO performance?

Use Search Query Performance if you have Brand Registry. Divide click share by impression share, cart-add share by click share and purchase share by cart-add share. The lowest ratio is the stage to fix. Without Brand Registry, track autocomplete phrases and your results-page position by hand.